We created this as a demo for OpenSearch on AWS (we are AWS partners) but I really enjoyed making it and I thought it turned out to be really fun.
Took the same type of dataset at the OP (I think we have 80,000 posters) but we used AWS’ multimodal embedding model to enable semantic search across the posters, as well as the ability to do a similarity search based on the posters.
Eventually we will include AWS’ personalize as part of the service to show off some of the movie recommendation engine features they developed.
Not commercialized there’s definitely some bugs, we built it for fun
We created this as a demo for OpenSearch on AWS (we are AWS partners) but I really enjoyed making it and I thought it turned out to be really fun.
Took the same type of dataset at the OP (I think we have 80,000 posters) but we used AWS’ multimodal embedding model to enable semantic search across the posters, as well as the ability to do a similarity search based on the posters.
Eventually we will include AWS’ personalize as part of the service to show off some of the movie recommendation engine features they developed.
Not commercialized there’s definitely some bugs, we built it for fun